A framework for user modeling in QuizMASter

Abstract

This paper presents an approach to user modeling in QuizMASter, a multi-user educational game shows that uses multi-agent systems to create a personalized learning environment. To keep the students motivated during the games, the system creates, maintains, and uses models of all attending contestants (i.e., learners). Taking advantage of prominent student modeling techniques with some novel ideas, we propose a structure and procedure to create student models and combined student models in QuizMASter. These models are used to create an adaptive environment in the quiz games, selecting and posing questions whose levels of difficulty closely match the knowledge levels of game contestants in certain areas of knowledge; meaning that the game questions for each game are adjusted to the knowledge levels of the learners and the game does not provide too easy or too difficult questions for the contestants. By selecting properly challenging questions for each game show, students are expected to stay motivated and continue benefiting from using the game show, increasing their knowledge through game-based learning.